Evolution of Segmentation: RollingBack the Clock

Dispatch

Andrew Peel and Michael Akam

Recent work has revealed striking similarities in thegenetic mechanisms underpinning somitogenesis inzebrafish and segmentation in the spider. Could thismean that the bilaterian common ancestor wassegmented after all?

When phylogenetic trees are constructed withmolecular data, the bilaterally symmetrical animals —the bilateria — fall into three distinct lineages [1]. It isstriking that segmented animals occur in each ofthese three branches, but in each case they aregrouped with others that are not segmented (Figure 1).This poses an intriguing question [2]. Was Urbilateria,the common ancestor of all bilaterian animals, seg-mented, with a metameric body plan that was secon-darily lost in many lineages?

The mechanisms used during somitogenesis invertebrates and segmentation in an annelid — the leech— and in an arthropod — the fruit fly Drosophila —exhibit gross differences, at both molecular and mor-phological levels. These differences have led many toconclude that metameric body plans evolved indepen-dently at least three times over the course of animalevolution. However, leeches and flies both exhibit evo-lutionarily ‘derived’ modes of segmentation, and themolecular mechanisms that underlie these may beunrepresentative of their phyla. In this context, recentwork by Stollewerk et al. [3] on the spider Cupienniussalei is of great interest, for spiders are chelicerates andthus represent a branch of the arthropods that divergedvery early from the insect/crustacean lineage [4].

The mechanism underlying segmentation of the che-licerate ‘abdomen’, the opisthosoma, appears to bedistinct from that operating in the anterior, appendage-bearing prosoma [5–6]. Opisthosomal segments formfrom a posterior growth zone, such that mature seg-ments appear one by one in an anterior to posteriorprogression. This is reminiscent of the way somitesarise from the anterior of the presomitic mesodermduring vertebrate somitogenesis [7].

Stollewerk et al. [3] have demonstrated that thesimilarity between vertebrate somitogenesis andopisthosomal segmentation is not merely morphologicalin nature. In both cases, the Notch signalling pathwayappears to play a key role. The authors found thatexpression of the gene delta-1, which encodes a ligandfor the spider receptor Notch, appears and then clearsrepeatedly from the posterior growth zone, in a fashionreminiscent of that previously reported for the spider

homologue of the Drosophila segmentation gene hairy[8]. This is striking, given that representatives of thesesame two gene families, deltaC and her-1/her-7, are alsoexpressed dynamically in the posterior presomiticmesoderm during zebrafish somitogenesis [9,10].

In zebrafish a wave of deltaC and her-1/her-7expression advances from posterior to anterior acrossthe presomitic mesoderm in tandem with the formationof each somite, a phenomenon which has been attrib-uted to the cycling on and off of these genes withinindividual cells, rather than cell migration [7,9,10].When cyclic deltaC or her-1/her-7 expression is dis-rupted, the somitic boundaries fail to form correctly.Dynamic expression of deltaC has been shown to bedependent on her-1/her-7 and vice versa [9,10]. Cyclicexpression is thought to be controlled by a ‘segmenta-tion clock’ or ‘oscillator’ [7]. Recently it has been pos-tulated that deltaC and her-1/her-7 are components ofthe oscillator itself [9,10] — which may explain theirapparent conservation — though recent results from astudy on mouse somitogenesis suggest their dynamicexpression may be an ‘output’ of the oscillator [11].

It is not yet clear whether the dynamic expression ofthe spider genes reflects oscillations in expressionwithin individual cells. Further similarities, however,

Current Biology, Vol. 13, R708–R710, September 16, 2003, ©2003 Elsevier Science Ltd. All rights reserved. DOI 10.1016/S0960-9822(03)00647-X

Laboratory for Development and Evolution, Department ofZoology, University of Cambridge, Downing Street,Cambridge CB2 3EJ, UK.

Figure 1. Three hypotheses for the evolution of segmentationduring bilaterian evolution.(1) A segmented bilaterian common ancestor (red dot) wouldrequire segmented body plans to have been lost at least threetimes, in lineages leading to deuterostome, ecdysozoan andlophotrochozoan phyla. (2) Segmentation could have arisentwice (blue dots), in the lineages leading to protostomes andvertebrates, respectively. This would require segmented bodyplans to have been lost at least twice in the lineages leading toecdysozoan and lophotrochozoan phyla. (3) The third hypothe-sis sees segmentation having arisen independently three times(black dots) in the lineages leading to vertebrates, annelids andarthropods. (Adapted from [2].)

Vertebrates

Annelids

Arthropods

Unsegmentedphyla

Unsegmented phyla

Unsegmented phyla

'Urbilateria'

Protostomecommonancestor

Deuterostomes Ecdysozoans Lophotrochozoans

Protostomes

Current Biology

suggest a zebrafish-like clock mechanism might wellbe operating during opisthosomal segmentation.Firstly, in both zebrafish and spider, the delta and hairyhomologues are expressed in overlapping domains[3,9,10]. Secondly, in both species, the knockdown ofgenes involved in Notch/Delta signalling leads to a dis-ruption in the dynamic expression of the hairy homo-logues and abnormal segment boundary formation[3,9,10]. Finally, the expression pattern of the singlespider notch gene is similar to the combined expres-sion patterns of two of the zebrafish notch homo-logues [3,12]: the uniform expression in the growthzone is reminiscent of notch1a expression throughoutthe presomitic mesoderm; and expression in bandsanterior to the growth zone may be equivalent to thesegmentally reiterated expression of notch5 in theanterior presomitic mesoderm — spider notch andzebrafish notch5 are expressed in the posterior offuture segments and somites, respectively.

In long-germ insects such as Drosophila, acascade of transcription factors acting in a cell-mem-brane-free environment — a syncytium — subdividesthe blastoderm in such a way that all segmentsdevelop simultaneously [13]. Notch signalling is notknown to be involved in this process. But segmenta-tion in short-germ insects, such as the grasshopperSchistocerca, resembles that of C. salei, with poste-rior segments appearing in much the same way asthose of the chelicerate opisthosoma [14].

One modulator of Notch signaling, Fringe, hasbeen examined in basal insects, such as thegrasshopper. In mouse [15] and chick [16] embryos,lunatic fringe is expressed in a dynamic fashion inthe posterior presomitic mesoderm, and is necessaryfor normal somitogenesis. This led Dearden et al. [17]

to examine the role of the Schistocerca fringe homo-logue during grasshopper segmentation: they foundthat fringe is expressed in a segmentally reiteratedpattern during embryogenesis, but downstream ofthe segmentation gene engrailed and too late to bepart of the primary segmentation process. On thisbasis, and because expression of notch itselfshowed no dynamic modulation during segmenta-tion, Dearden et al. [17] concluded that Notch signal-ing was unlikely to play a part in the formation ofgrasshopper segments.

A role for Notch signalling in the segmentation oflower insects cannot, however, be ruled out. Inhindsight, the choice of fringe as a marker for Notch-dependent segmentation appears to have beenunfortunate. In zebrafish, lunatic fringe is not involvedin somitogenesis [18]. Rather, it is expressed in asegmentally reiterated pattern downstream of somi-togenesis, suggesting a role similar to its Schisto-cerca homologue. Thus, lunatic fringe appears tohave been recruited to play a role in somitogenesis inhigher vertebrates, and so its homologue would notnecessarily be expected to feature in an arthropodclock mechanism. The expression patterns of hairyand delta homologues in basal insects should provemore informative.

It is tempting to speculate that different mechanismsfor making anterior and posterior segments are anancestral feature of arthropod development [19], withthe posterior mechanism involving a zebrafish-likeclock. This idea is attractive, because it allows us toenvisage how the transition from short-germ to long-germ modes of development might have occurredduring insect evolution. The number of segments pat-terned by a clock might have been progressively

Current BiologyR709

Figure 2. A model for the transition fromshort germ to long germ modes of seg-mentation during insect evolution.The number of segments patterned in agrowth zone by a putative segmentationclock (blue) decreases as segmentationgenes downstream of the clock comeunder the control of newly recruited gapgenes. Coloured blocks represent newlyevolved gap gene response elements,similar to those found in the Drosophilahairy and even-skipped [20] genes. It isunclear as to whether the change to pat-terning in a cell-membrane-free (syncytial)environment would be a prerequisite forthis transition, or could have occurred inparallel with it.

Current Biology

Evolutionary tim

e

Cel

lula

rS

yncy

tial

Embryonic development

Germband Segmented embryo Pair-rule cis-regulation

SegmentationClock

Segmentationclock

Segmentationclock

Proto-gapgene

Proto-gapgenes

Gap genes

Short germ ancestor

Long germ insect

DispatchR710

reduced during evolution, with mechanisms analogousto those already operating in the anterior taking over.

A prediction of this model (Figure 2) might be thatthe primary pair-rule genes of Drosophila ancestrallyfunctioned downstream of a clock. This would explainhow the complex cis-regulatory elements, controllingstriped expression of these genes [20], evolved — onestripe enhancer may have evolved at a time as theinfluence of the clock on anterior segments retreatedand anterior determinants (the future gap genes) tookover. Indeed the primary pair-rule genes even-skippedand runt appear to be expressed in a reiterated patternin the opisthosoma of the spider [8].

A clock-like mechanism in arthropod segmentationremains to be demonstrated. Even if one proves to beinvolved, it would be premature to conclude that Urbila-teria was segmented, for Notch-mediated patterning isused in many and diverse biological processes. Thesimilarities between vertebrate somitogenesis andarthropod segmentation might still be analogous, nothomologous. To test whether a segmentation clockwas an ancestral feature of bilaterians, we must lookmore closely at the genes and gene interactionsinvolved, and must survey diverse representatives ofbasally branching clades within the arthropods,annelids and chordates, as well as less obviously seg-mented phyla, such as the molluscs. Clearly, this willtake time. Despite recent advances the debate as towhether Urbiliteria was segmented looks set to con-tinue for some time to come.

References1. Aguinaldo, A.M., Turbeville, J.M., Linford, L.S., Rivera, M.C., Garey,

J.R., Raff, R.A. and Lake, J.A. (1997). Evidence for a clade of nema-todes, arthropods and other moulting animals. Nature 387, 489-493.

2. Davis, G.K. and Patel, N.H. (1999). The origin and evolution of seg-mentation. Trends Genet. 15, M68-M72.

3. Stollewerk, A., Schoppmeier, M. and Damen, W.G.M. (2003).Involvement of Notch and Delta genes in spider segmentation.Nature 423, 863-865.

4. Cook, C.E., Louise Smith, M., Telford, M.J., Bastianello, A. andAkam, M. (2001). Hox genes and the phylogeny of the arthropods.Curr. Biol. 11, 759-763.

5. Damen, W.G.M. (2002). Parasegmental organization of the spiderembryo implies that the parasegment is an evolutionary conservedentity in arthropod embryogenesis. Development 129, 1239-1250.

6. Dearden, P.K., Donly, C., and Grbic, M. (2002). Expression of pair-rule gene homologues in a chelicerate: early patterning of the two-spotted spider mite Tetranychus urticae. Development 129,5461-5472.

7. Saga, Y. and Takeda, H. (2001). The making of the somite: molecu-lar events in vertebrate segmentation. Nat. Rev. Genet. 2, 835-845.

8. Damen, W.G.M., Weller, M. and Tautz, D. (2000). Expression pat-terns of hairy, even-skipped, and runt in the spider Cupiennius saleiimply that these genes were segmentation genes in a basal arthro-pod. Proc. Natl. Acad. Sci. USA 97, 4515-4519.

9. Holley, S.A., Julich, D., Rauch, G-J., Geisler, R. and Nusslein-Volhard, C. (2002). her1 and the notch pathway function within theoscillator mechanism that regulates zebrafish somitogenesis.Development 129, 1175-1183.

10. Oates, A.C. and Ho, R.K. (2002). Hairy/E(spl)-related (Her) genes arecentral components of the segmentation oscillator and displayredundancy with the Delta/Notch signaling pathway in the forma-tion of anterior segmental boundaries in the zebrafish. Development129, 2929-2946.

11. Aulehla, A., Wehrle, C., Brand-saberi, B., Kemler, R., Gossler, A.,Kanzler, B. and Herrmann, B.G. (2003). Wnt3a plays a major role inthe segmentation clock controlling somitogenesis. Dev. Cell 4, 385-406.

12. Westin, J. and Lardelli, M. (1997). Three novel Notch genes inzebrafish: implications for vertebrate Notch gene evolution andfunction. Dev. Genes Evol. 207, 51-63.

13. St Johnston, D. and Nusslein-Volhard, C. (1992). The origin ofpattern and polarity in the Drosophila embryo. Cell 68, 201-219.

14. Bentley, D., Keshishian, H., Shankland, M. and Toroian-Raymond,A. (1979). Quantitative staging of embryonic development of thegrasshopper, Schistocerca nitens. J. Embryol. Exp. Morph. 54, 47-74.

15. Forsberg, H., Crozet, F., and Brown, N.A. (1998). Waves of mouseLunatic fringe expression, in four-hour cycles at two-hour intervals,precede somite boundary formation. Curr. Biol. 8, 1027-1030.

16. McGrew, M.J., Dale, J.K., Fraboulet, S. and Pourquie, O. (1998). Thelunatic Fringe gene is a target of the molecular clock linked tosomite segmentation in avian embryos. Curr. Biol. 8, 979-982.

17. Dearden, P. and Akam, M. (2000). A role for Fringe in segment mor-phogenesis but not segment formation in the grasshopper, Schis-tocerca gregaria. Dev. Genes. Evol. 210, 329-336.

18. Leve, C., Gajewski, M., Rohr, K.B. and Tautz, D. (2001). Homologuesof c-hairy1 (her9) and lunatic fringe in zebrafish are expressed in thedeveloping central nervous system, but not in the presomitic meso-derm. Dev. Genes Evol. 211, 493-500.

19. Dearden, P. and Akam, M. (2001). Early embryo patterning in thegrasshopper, Schistocerca gregaria: wingless, decapentaplegic andcaudal expression. Development 128, 3435-3444.

20. Fujioka, M., Emi-Sarker, Y., Yusibova, G.L., Goto, T. and Jaynes,J.B. (1999). Analysis of an even-skipped rescue transgene revealsboth composite and discrete neuronal and early blastodermenhancers, and a multi-stripe positioning by gap gene repressorgradients. Development 126, 2527-2538.